Fix max_seqlets_subsample bug

Mirror the fix in: jmschrei/tfmodisco-lite#60

Author: defunkt

examples/MergeMotifsAcrossRuns.ipynb

@@ -628,7 +628,7 @@
         "min_num = 30 # also called min_num_to_trim_to\n",
         "flank_to_add = 5 # also called initial_flank_to_add\n",
         "window_size = 20 # also called trim_to_window_size\n",
-        "max_seqlets_subsample = 300 # also called merging_max_seqlets_subsample\n",
+        "max_seqlets_subsample = 1000 # also called merging_max_seqlets_subsample\n",
         "\n",
         "pattern_group = 'pos_patterns' #pos_patterns or neg_patterns\n",
         "\n",

modiscolite/aggregator.py

@@ -186,7 +186,7 @@ def _detect_spurious_merging(patterns, track_set, perplexity,
 				prob_and_pertrack_sim_merge_thresholds=prob_and_pertrack_sim_merge_thresholds,
 				prob_and_pertrack_sim_dealbreaker_thresholds=prob_and_pertrack_sim_dealbreaker_thresholds,
 				min_frac=min_frac, min_num=min_num, flank_to_add=flank_to_add, window_size=window_size, 
-				bg_freq=bg_freq, max_seqlets_subsample=1000)
+				bg_freq=bg_freq, max_seqlets_subsample=max_seqlets_subsample)
 
 			to_return.extend(refined_subpatterns[0]) 
 		else:
@@ -197,7 +197,7 @@ def _detect_spurious_merging(patterns, track_set, perplexity,
 				prob_and_pertrack_sim_merge_thresholds=prob_and_pertrack_sim_merge_thresholds,
 				prob_and_pertrack_sim_dealbreaker_thresholds=prob_and_pertrack_sim_dealbreaker_thresholds,
 				min_frac=min_frac, min_num=min_num, flank_to_add=flank_to_add, window_size=window_size, 
-				bg_freq=bg_freq, max_seqlets_subsample=1000)
+				bg_freq=bg_freq, max_seqlets_subsample=max_seqlets_subsample)
 
 def SimilarPatternsCollapser(patterns, track_set,
 	min_overlap, prob_and_pertrack_sim_merge_thresholds,

modiscolite/tfmodisco.py

@@ -7,7 +7,6 @@
 import scipy
 import scipy.sparse
 
-from collections import OrderedDict
 from collections import defaultdict
 
 from . import affinitymat
@@ -161,7 +160,7 @@ def seqlets_to_patterns(seqlets, track_set, track_signs=None,
 	final_flank_to_add=0,
 	prob_and_pertrack_sim_merge_thresholds=[(0.8,0.8), (0.5, 0.85), (0.2, 0.9)],
 	prob_and_pertrack_sim_dealbreaker_thresholds=[(0.4, 0.75), (0.2,0.8), (0.1, 0.85), (0.0,0.9)],
-	subcluster_perplexity=50, merging_max_seqlets_subsample=300,
+	subcluster_perplexity=50, merging_max_seqlets_subsample=1000,
 	final_min_cluster_size=20,min_ic_in_window=0.6, min_ic_windowsize=6,
 	ppm_pseudocount=0.001):
 
@@ -275,7 +274,7 @@ def TFMoDISco(one_hot, hypothetical_contribs, sliding_window_size=21,
 	initial_flank_to_add=10, final_flank_to_add=0,
 	prob_and_pertrack_sim_merge_thresholds=[(0.8,0.8), (0.5, 0.85), (0.2, 0.9)],
 	prob_and_pertrack_sim_dealbreaker_thresholds=[(0.4, 0.75), (0.2,0.8), (0.1, 0.85), (0.0,0.9)],
-	subcluster_perplexity=50, merging_max_seqlets_subsample=300,
+	subcluster_perplexity=50, merging_max_seqlets_subsample=1000,
 	final_min_cluster_size=20, min_ic_in_window=0.6, min_ic_windowsize=6,
 	ppm_pseudocount=0.001, verbose=False):